Beruflich Dokumente
Kultur Dokumente
Users Manual
(Version 1.2 for DOS)
Asphalt
Base
Roadbed
Ronald S. Harichandran
Gilbert Y. Baladi
by
Ronald S. Harichandran
Associate Professor
and
Gilbert Y. Baladi
Professor
January 2000
ii
Table of Contents
1.
2.
3.
4.
5.
6.
6.
Introduction.................................................................................................................1
Summary of Modeling and Analysis ..........................................................................1
2.1
Modeling of the Pavement ........................................................................1
2.2
Granular and Cohesive Material Models ..................................................1
2.3
Gravity and Lateral Stresses .....................................................................1
2.4
Finite Element Analysis............................................................................2
2.5
Computation of Stresses and Strains at Layer Interfaces..........................4
2.6
Estimated Equivalent Resilient Moduli ....................................................4
2.7
Fatigue and Rut Depth Prediction.............................................................4
System Requirements..................................................................................................4
Configuring the Computer ..........................................................................................5
4.1
Installation Procedure ...............................................................................5
4.2
The CONFIG.SYS File.............................................................................5
4.3
Required Amount of Free Memory...........................................................5
4.4
Printing Graphics ......................................................................................5
4.5
Running MICHPAVE for the First Time..................................................6
Using MICHPAVE .....................................................................................................6
5.1
Filenames ..................................................................................................6
5.2
Cursor Movement and Editing Keys.........................................................6
5.3
Title Screen ...............................................................................................7
5.4
Main Menu................................................................................................7
5.5
Data File Menus and Associated Data-Entry Forms.................................9
5.6
Performing Analysis ...............................................................................16
5.7
Plotting the Results .................................................................................18
5.8
Printing the Results .................................................................................20
Problem Reporting ....................................................................................................20
References.................................................................................................................20
iii
List of Figures
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.
Figure 21.
Figure 22.
Figure 23.
Figure 24.
List of Tables
Table 1.
iv
1. Introduction
MICHPAVE is a user-friendly, non-linear finite element program for the analysis of flexible
pavements. The program computes displacements, stresses and strains within the pavement due to
a single circular wheel load. Useful design information such as fatigue life and rut depth are also
estimated through empirical equations.
Most of MICHPAVE is written in FORTRAN 77. Graphics and screen manipulations are
performed using the FORTRAN callable GRAFMATIC graphics library, marketed by Microcompatibles Inc., 301 Prelude Drive, Silver Spring, MD 20901.
when ( 1 3 ) K 1
when ( 1 3 ) > K 1
log MR
K2
1
log K1
log
Figure 1
h = K0g
where K0 = coefficient of earth pressure at rest. For granular soils K0 = 1 sin and for cohesive
soils K0 = 1 0.95 sin , where = angle of internal friction.
To approximately account for locked-in stresses caused by compaction, the user can input
a value for K0 higher than the coefficient of earth pressure at rest.
2.4 Finite Element Analysis
Rectangular four-noded axisymmetric finite elements with linear interpolation functions are
used in all upper layers and through the depth specified by the user for the last layer (the roadbed).
A lateral boundary is placed at a radial distance of 10a from the center of the loaded area, where
a = radius of the loaded area. A default mesh is initially generated, but this may be modified by the
user. The default mesh has the following characteristics:
MR
K3
1
K4
K2
1
1 3
K1
Figure 2
In the radial direction, the total width of 10 radii is divided into four regions. Within any
region, all elements have the same horizontal dimension. The first region, between 0 and 1
radius, is equally divided into four elements; the second region, between 1 radius and 3 radii, is equally divided into four elements; the third region, between 3 radii and 6 radii, is
equally divided into three elements; and the fourth region, between 6 radii and 10 radii, is
equally divided into two elements.
Within any layer, all elements have the same vertical dimension. The number of elements
in each layer in the vertical direction is dependent on the layer thickness, but at least four
elements are used in the top (AC) layer, and at least two elements are used in all other layers.
A typical default finite element mesh is shown in Fig. 3.
Displacements, stresses and strains are computed only within the region modeled by finite
elements. In order to increase accuracy, and to reduce the memory and computation time required
by the program, the infinite extent of the last layer is modeled by using a flexible bottom boundary
(Harichandran and Yeh 1989). The half-space below the bottom boundary is assumed to be homogeneous and linear elastic. The modulus of the half-space is taken as the average moduli of the finite elements immediately above the bottom boundary.
The non-linear analysis consists of several iterations. A linear analysis is performed in each
iteration, after which the resilient modulus of each finite element is revised if necessary. If the
Mohr-Coulomb failure criterion is violated in any granular or cohesive soil element, then the principal stresses are modified to reflect the failure condition, and the resilient moduli are determined
from the modified stresses (Raad and Figueroa 1980). The iteration is repeated until the resilient
moduli of all the elements stabilize.
Depth
0.0"
Asphalt
10.0"
Base
30.0"
Roadbed
3a
6a
50.0"
10 a
3. System Requirements
The MICHPAVE program was originally written for IBM compatible personal computers
running under DOS. Presently it is also available for Sun and Hewlett-Packard workstations running under UNIX. For DOS systems, the following hardware and software are required:
PC-DOS or MS-DOS version 3.0 or higher
640 KB of random access memory (RAM)
A hard disk
A color graphics adapter (CGA, EGA or VGA) and compatible monitor
Although not strictly required for the use of the program, the following hardware is strongly
recommended:
A math co-processor (8087, 80287 or 80387). Running time will be greatly increased if a
math co-processor is not installed.
A printer for obtaining hardcopies of plots and output.
then this statement will need to be removed and the computer re-booted (by simultaneously pressing the CTRL, ALT and DEL keys) before running MICHPAVE. If available, use of an ANSI.SYS
compatible device driver that can be unloaded from memory on demand is convenient since it eliminates the need to re-boot the computer.
4.3 Required Amount of Free Memory
The MICHPAVE program requires about 515 KB of free memory to run. DOS and memory
resident programs (such as SIDEKICK) reduce the amount of free memory for use by other programs. The amount of free memory available can be checked by using the DOS command
CHKDSK. If there is insufficient free memory, then memory resident programs will need to be removed before running MICHPAVE.
If there is insufficient memory to load the program, the following message will be displayed:
Program too big to fit in memory.
Sometimes the program may load into memory without any problem, but the following error message may be displayed during computations:
Run-time error F6700:
-heap space limit exceeded.
it is issued every time the computer is turned on. To download graphics that are on the screen to
the printer simply press the SHIFT and PrScr keys simultaneously.
4.5 Running MICHPAVE for the First Time
To run MICHPAVE simply type: MICHPAVE. When running for the first time, the program
will request the following information about the computer system:
Which graphics adapter and monitor do you have (MONO/CGA/EGA)?
Is your computer strictly IBM compatible (Y/N)?
Is your printer EPSON or EPSON compatible (Y/N)?
The response to the above prompts are stored in a file named SYSTEM.DAT. When running
MICHPAVE subsequently, the system information is read from this file. In case a mistake is made
when specifying this information, or if the graphics adapter in the computer or the printer is
changed at a later time, the file SYSTEM.DAT should be deleted before running MICHPAVE so
that it will prompt again for a description of the new hardware.
The graphics resolution for EGA systems will be substantially higher than for CGA systems.
For VGA systems, specify EGA.
If the computer is not strictly IBM compatible, then problems may be encountered with the
data-entry forms due to incompatibility with the graphics software, if the computer had originally
been specified as being fully IBM compatible. By defining the computer to be not strictly IBM
compatible MICHPAVE can still be used, but some of the color used to enhance the data-entry
forms will be lost.
For EPSON compatible printers MICHPAVE automatically sets the print mode to condensed
when printing the output after an analysis, so that the 132-column wide output file is printed properly. If the printer is not EPSON compatible, then its print mode will need to be set externally before printing the output. For an EPSON printer with a wide carriage capable of printing 132
characters per line in normal mode, specify the printer to be non-EPSON compatible.
5. Using MICHPAVE
MICHPAVE is designed to be user-friendly. Menus are used to perform the required steps in
pavement analysis, and data-entry forms facilitate data input. In addition, extensive checking of input data is performed and appropriate error messages are displayed upon completion of each dataentry form.
5.1 Filenames
The names of files in which the data and results are saved may include a pathname if necessary (e.g., A:I-96.DAT to save the file I-96.DAT on the diskette in drive A:, \JOB1\I-96.DAT to
save the file in subdirectory JOB1, etc.). If no path is specified, the file will be saved in the default
subdirectory.
5.2 Cursor Movement and Editing Keys
The data-entry forms have several fields into which data is typed. The field in which the cursor resides is highlighted on IBM compatible systems. The functions of the cursor movement and
editing keys within a data-entry form are described in Table 1.
TABLE 1
KEY
Return/Enter
Tab
Shift Tab
Home
End
or PgUp
or PgDn
Backspace
Del
Ins
F1
Esc
Version 1.2
MICHPAVE
Nonlinear Finite Element Program
for Analysis of Flexible Pavements
Developed for
Michigan Department of Transportation
by
Dept. of Civil & Environmental Engineering
Michigan State University
East Lansing, MI 48824-1226
Title screen
Credits screen
MAIN MENU
1.
2.
3.
4.
5.
6.
7.
8.
9.
Overview
Create a New Data File
Change Current Data File
Modify an Existing Data File
Perform Analysis
Type summary results on screen
Plot Results on screen
Print Results on printer
Exit-Return to DOS
Selection:
Figure 6
Main menu
Option 8: Used after analysis to print results on the printer from within the MICHPAVE
program. The output requires a line width of 132 characters. EPSON compatible printers
are automatically set to condensed mode by the program.
Option 9: Terminates the MICHPAVE program and returns to DOS.
5.5 Data File Menus and Associated Data-Entry Forms
Data File, Modify Current Data File, and Modify Existing Data File menus are displayed
when selecting options 2, 3 or 4, respectively, from the main menu. All three menus are identical
in structure, and the first is shown in Fig. 8. The only difference between these menus is that in
MAIN MENU
SUBMENUS
1.Overview
1.Initial Data
2.Layer Type
1.Asphalt
3.Material Properties
5.Perform Analysis
2.Granular
3.Cohesive
Cross
Section
Locations
2.
3.
4.
5.
6.
7.
Selection:
Figure 8
those used for modifying data files, existing data is modified instead of specifying new data. It is
recommended that the options in the data file menu be followed in sequence.
When modifying an existing file it is mandatory to first use option 1 and specify new names
for the required filenames, or to indicate that the input and output files used earlier should be overwritten. The other options may be performed in any sequence.
5.5.1 Option 1: Initial, fatigue life and rut depth data
This option displays the data-entry form shown in Fig. 9 (typical data is also shown in bold
typeface).
INITIAL DATA
1.
2.
3.
4.
Number of Layers: 3
5.
Wheel Load:
6.
7.
(max. 6)
9000.0 (lb.)
Figure 9
(psi)
10
The data that should be entered into the fields are described below:
1. Filename to Save Data to: All data that is entered in this and other forms is stored in this
file. The data may be recovered at a later time and modified if necessary.
2. Filename to Output Results to: The output from the analysis will be directed to this file.
The output is in standard ASCII form and may be viewed or edited using any text editor. It
should be noted however that a line width of 132 characters is used for the output.
3. Title: This is a description of the current job for identification purposes.
4. Number of Layers: The number of layers in the pavement section. A maximum of six layers are permitted. Note that the roadbed soil (subgrade) is counted as one layer.
5. Wheel Load: Equal to half the axle load in pounds.
6. Tire Pressure: The pressure in the truck tire in psi.
7. Fatigue Life & Rut Depth Computation Required (Y/N)? The user should respond with
a Y if fatigue life and rut depth in the section are to be estimated. Empirical expressions are
used to relate the fatigue life and rut depth to results from the mechanistic analysis. These
relations are currently valid only for three-layer pavements with AC, base and roadbed soil
layers, and for four-layer pavements with AC, base, subbase and roadbed soil layers (Baladi 1989). Fatigue life and rut depth estimates for other pavement sections may not be
meaningful. The rut depth is estimated for the number of load repetitions causing fatigue
failure of the pavement.
Answering in the affirmative to question 7 in the Initial Data form displays the data-entry
form shown in Fig. 10, which is used to enter data for the fatigue life and rut depth calculations.
The data table below the form shows typical kinematic viscosities for different asphalt grades.
The data that should be entered into the fatigue life and rut depth form are described below:
1. Average Annual Temperature: The average annual air temperature expected at the pavement location.
2.
3.
Kinematic Viscosity:
Asphalt Grade
77.00
(Fahrenheit)
3.00
270.00 (centistoke)
AC 2.5
AC 5.0
AC 10
159
212
270
Figure 10
11
LAYER TYPE
1 Asphalt or Linear; 2 Granular; 3 Cohesive
Layer number (from top)
Type (1,2, or 3)
3
Figure 11
2. Percent Air Voids in Asphalt Mix: The percent air voids in the asphalt mix as expected
in the field.
3. Kinematic Viscosity: The kinematic viscosity of the asphalt binder.
5.5.2 Option 2: Layer type
The type of material used for each layer in the pavement section is identified by typing 1, 2
or 3 for asphalt, granular or cohesive soil layers, respectively, into the form shown in Fig. 11. Asphalt is treated as a linear elastic material in the analysis. Lime asphalt or cement treated materials
may be specified as type 3. To perform a linear analysis of the entire pavement section specify all
layers to be of type 1. Specifying types 2 or 3 implies non-linear analysis.
5.5.3 Option 3: Material properties
Three different sets of material properties are required for the three material types 1, 2 and 3.
Properties for layers with asphalt or linear elastic materials, including the names of the layers,
thicknesses, resilient moduli, Poisson's ratios, densities, and coefficients of lateral earth pressure
(K0), are specified in the form shown in Fig. 12. For compacted layers, the locked-in lateral
stresses can be approximately accounted for by specifying a relatively large value for K0 (e.g., larger than 0.4).
Name of Layer
Asphalt
Thickness
(inches)
Modulus
(psi)
10.0
500000.0
Poisson's
Density
Ratio
(lb/cu.ft)
.35
150.0
Ko
.40
12
20.0
Ko
K1
(psi)
K2
.40
9000.0
Cohesion
Density
(psf)
(degree) (pcf)
.35 .40
.0
30.0
120.0
K1
Silty Sand
Sand Gravel
Crushed Gravel
Slag
1620
4480
7210
24250
K2
Material Type
K1
.62
.53
.45
.37
Sand/Aggregate
4350
Partially Crushed Gravel 5967
Limerock
14030
K2
.59
.52
.40
Properties for granular layers, including the names of the layers, thicknesses, coefficients of
lateral earth pressure (K0), K1 and K2 parameters, Poisson's ratios (), cohesions, friction angles
(), and densities are specified in the form shown in Fig. 13. Typical values of the parameters K1
and K2 for a variety of granular soils are displayed in the table below the form.
Properties for cohesive layers, including the names of the layers, thicknesses, coefficients of
lateral earth pressure (K0), K1, K2, K3 and K4 parameters, Poissons ratios (), cohesions, friction
angles (), and densities are specified in the form shown in Fig. 14. Typical values of the parame-
Name of Thick. Ko K1
K2
Layer
(in.)
(psi) (psi)
Roadbed
20.0 .40
6.0
K3
3020.0
1110.0
K4
178. .45
Coh.
(psf) (deg)
800.0
.0
Dens.
(pcf)
120.0
13
Figure 15
Data-entry form for specifying the number of cross sections along which results are computed
ters K1, K2, K3 and K4 are given in the notes. These parameters have currently not been established
widely for different cohesive soils.
It should be noted that the thickness specified for the last layer (roadbed soil) controls the
depth to which displacements, stresses, and strains are computed. A thickness of about 6" to 12" is
recommended. For analysis, the last layer is actually considered to be semi-infinite.
5.5.4 Option 4: Cross sections for computation of results
Displacements, stresses and strains are computed along horizontal and vertical cross sections
specified by the user. The number of horizontal and vertical sections are specified in the data-entry
form shown in Fig. 15. At least one vertical section must be used.
The depths at which the horizontal sections are located are specified in the data-entry form
shown in Fig. 16. To aid in these specifications the thickness of each layer in the pavement section
is displayed in the upper table on the right. Although the horizontal sections may be specified at
any depth within the pavement, in the finite element method some stresses and strains are most accurately computed at the center of elements. Thus, best results will be obtained if the locations of
the horizontal sections correspond to the mid-depths of elements. Optimal locations within each
layer, corresponding to the mid-depths of the elements in that layer, are shown in the lower table
Section No.
Depth (inches)
.00
10.00
28.00
33.30
Figure 16
Layer
Name
Thick (in.)
1.
2.
3.
Asphalt
Base
Subgrade
10.0
20.0
20.0
1.3,
3.8,
6.3,
8.8
2.
3.
14
on the right. Nevertheless, it is strongly recommended that horizontal sections be specified at each
layer interface. Note that the most critical stresses are compression at the top and tension at the bottom of the AC surface, and compression at the top of the roadbed soil.
The radial distances at which the vertical sections are located are specified in the data-entry
form shown in Fig. 17. The first section must be located at the center of the loaded area (r = 0").
Although the other vertical sections may be specified at any radial distance within the pavement
modeled by finite elements (0 to 10 radii of the loaded area), some stresses and strains are most
accurately computed at the center of elements. Thus, best results will be obtained if the locations
of the vertical sections correspond to the middle of an element. In MICHPAVE, the elements are
grouped into three regions in the radial direction, from 0 to a, a to 3a, 3a to 6a, and 6a to 10a, where
a = radius of loaded area. Optimal radial locations, corresponding to the middle of the elements,
are shown in the table on the right. Due to edge effects of the right boundary, it is recommended
that vertical section not be specified in the last region from 6a to 10a. The radius of the loaded area
is shown in the note below the tables.
5.5.5 Option 5: Plot finite element mesh
This option plots the current finite element mesh on the screen. The loaded region and the
radius of the loaded area, a, are also shown. (Fig. 3 shows a typical finite element mesh for the
mesh parameters given in bold typeface in Figs. 18 and 19.)
5.5.6 Option 6: Modify finite element mesh
This option is used to modify the current finite element mesh. MICHPAVE automatically
generates a default mesh that should be sufficient for most purposes. However, for greater accuracy, or for unusual situations, the user may wish to modify the default mesh. Memory limitations
may, however, preclude the use of a very fine (large) mesh. First, the data-entry form shown in Fig.
18 is displayed for modifying the number of elements in the vertical direction, and the current number of elements within each layer are shown. All elements within a given layer have the same vertical dimension. For the default number of elements in the horizontal direction (13), the maximum
number of elements in the vertical direction are currently limited to 24.
Section
1
2
Rad. Dist.(in.)
.00
4.7
.7,
a - 3a
6.7,
3a - 6a
2.0,
3.3,
15
4.7
Layer
Thickness
Number of Elements
1.
Asphalt
10.0
2.
Base
20.0
3.
Subgrade
20.0
Total vertical elements 24; when total hori. elements = 13 (default value).
Figure 18 Data-entry form for modifying the number of elements in the vertical direction
After the required changes are made, the data-entry form shown in Fig. 19 is displayed for
modifying the number of elements in the horizontal direction, and the current number of elements
in the ranges 0 to a, a to 3a, 3a to 6a, and 6a to 10a, where a = radius of loaded area, are shown.
All elements within a given range have the same horizontal dimension.
5.6 Performing Analysis
The analysis portion of MICHPAVE consists of an initialization part, several iterations (for
non-linear material), and a concluding part. The number of iterations required for convergence of
the non-linear solution depends on the properties of the pavement section being analyzed, and on
the magnitude of the wheel load. Weaker sections will in general require a larger number of iterations for convergence. The maximum number of iterations allowed is 25. Pavements requiring
more iterations than this will probably be too weak to be practicable. The stage of analysis and the
time required for the previous stages are displayed on the screen during computation. A typical dis-
Number of Elements
1. R =
0 -
2. R =
a -
3a
3. R = 3a -
6a
4. R = 6a - 10a
5.35 inches
Figure 19 Data-entry form for modifying the number of elements in the horizontal direction
16
0 min 24 sec)
Iteration 1
(Completion time =
0 min 50 sec)
Iteration 2
(Completion time =
1 min
Conclusion
(Completion time =
0 min 24 sec)
2 sec)
2 min 40 sec
play is shown in Fig. 20 (the times shown were obtained on a PC with 80286 and 80287 processors).
After the analysis is completed a design summary is displayed, showing key design information such as the maximum tensile strain at the bottom of the asphalt layer, the average compressive
strain in the asphalt layer, the maximum compressive strain at the top of the roadbed soil, the number of equivalent standard axle loads required to cause fatigue failure, and the rut depth at the fatigue life. A typical summary is shown in Fig. 21. The caution statement at the bottom of the table
is a warning that if the estimated fatigue life is greater than 20 million load repetitions, then failure
would most probably occur due to thermal cracking rather than fatigue. The implication here is that
20 million ESAL will span a period of greater than 15 to 20 years. Hence asphalt hardening and
block cracking should be considered.
DESIGN SUMMARY
1. Max. Tensile strain in the asphalt layer =
1.116e-04
8.947e-05
1.112e-04
1.204e+08 ESAL
1.885e-01 (in)
6.966e-02 (in)
7. Expected rut depth in the base and/or subbase course = 9.097e-02 (in)
8. Expected rut depth in the roadbed soil =
2.791e-02 (in)
17
2.
3.
Selection:
Figure 22
These questions allow the user to output the summary results to the printer, and to recompute new
fatigue life and rut depth estimates for different values of annual temperature, and kinematic viscosity of the asphalt binder. Answering in the affirmative to the second question displays the Fatigue Life and Rut Depth data-entry form (see Fig. 10) on which these input data may be changed.
Note that the re-estimation of the fatigue life and rut depth for changes in this data is done using
empirical equations, and does not require a re-analysis. Also note that this is the only stage at which
fatigue life and rut depth may be estimated for the pavement for new input data, without performing a re-analysis. If the fatigue life and rut depth are not recomputed at this stage, but are desired
at a later time for the same pavement section, then the analysis will need to be performed again.
All calculations of the fatigue life and rut depth will be saved in the output file.
5.7 Plotting the Results
After an analysis, the results may be plotted. When this option is chosen before an analysis,
the results from the previous analysis will be plotted provided the files V.PLT and R.PLT which
contain the data for plots have not been deleted. Every analysis overwrites these plot files, so that
only one set is maintained at any given time. Results may be plotted along the vertical and horizontal sections previously specified by the user by selecting from the menu shown in Fig. 22.
The menu in Fig. 23 is used to select the quantities that may be plotted along vertical sections.
The vertical compressive and radial tensile stresses and the radial tensile strains are the quantities
that are commonly plotted. These are grouped together, and other quantities that may be plotted
are grouped below them.
The menu in Fig. 24 is used to select the quantities that may be plotted along horizontal sections. The vertical compressive stresses and the vertical deflections are the quantities that are commonly plotted. These are grouped together, and other quantities that may be plotted are grouped
below them.
18
2.
3.
4.
5.
Vertical deflections
6.
Radial displacements
7.
Shear stresses
8.
Tangential stresses
9.
Selection:
Figure 23 Menu for plots along vertical cross sections
2.
Vertical deflections
3.
4.
5.
6.
Radial displacements
7.
Shear stresses
8.
Tangential stresses
9.
Selection:
Figure 24
19
6. Problem Reporting
Although MICHPAVE has been tested quite extensively, it is possible that errors causing the
program to terminate abnormally may still be encountered if a haphazard sequence of options is
used. To report a problem, note down the number and message displayed when the program terminates abnormally, and send it along with a diskette containing the input data file to:
Dr. Ronald Harichandran or Dr. Gilbert Baladi
Department of Civil & Environmental Engineering
Michigan State University
East Lansing, MI 48824-1226
Alternatively, report the problem by e-mail to harichan@egr.msu.edu or baladi@egr.msu.edu.
References
Baladi, G.Y. (1989). Fatigue life and permanent deformation characteristics of asphalt concrete mixes, Transportation Research Record, 1227, 7586.
Harichandran, R.S., Baladi, G. Y., and Yeh, M-S. (1989). Development of a computer program for design of pavement systems consisting of layers of bound and unbound materials, Report No. FHWA-MI-RD-89-02, Michigan Department of Transportation, Lansing, Michigan.
Harichandran, R.S. and Yeh, M-S. (1989). Flexible boundary in finite element analysis of pavements, Transportation Research Record, 1207, 5060.
Harichandran, R. S., Yeh, M-S., and Baladi, G. Y. (1990). MICH-PAVE: A nonlinear finite element program for the
analysis of flexible pavements. Transportation Research Record, 1286, 123131.
Raad, L., and Figueroa, J. L. (1980). Load response of transportation support system, Journal of Transportation Engineering, ASCE, 106, 111128.
Yeh, M-S. (1989). Nonlinear finite element analysis of flexible pavements, dissertation submitted in partial fulfillment of the degree of Doctor of Philosophy, Michigan State University, East Lansing, Michigan.
20